Data-to-text (D2T) systems generate understandable textual reports from data. This discipline is currently experiencing a bursting scientific, technical and commercial expansion due to the rise of the Big Data era. D2T solutions help analysts, experts and users in general in saving time by performing data analysis and delivering relevant information as high quality texts. Likewise, the use of fuzzy sets and fuzzy logic as tools for obtaining meaningful linguistic information from data which also supports uncertainty management has allowed in recent times the emergence of an extensive research work focused on the generation of what are commonly known as “linguistic descriptions of data”. In this context, this PhD thesis provides a convergence point for both disciplines by studying how fuzzy sets can be applied to D2T systems in order to model the vagueness inherent to human language in the terms and expressions to be conveyed.
Keywords: data-to-text, fuzzy sets, fuzzy logic, natural language generation